Intelligent Home Hospital Scheduling System
Herath Mudiyanselage, Sampath Sujeewa Herath; Rathnayake Pathiranage, Nadeesha Subhashini Dilhani (2026)
2026
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202602112572
https://urn.fi/URN:NBN:fi:amk-202602112572
Tiivistelmä
Home Hospital Care (HaH) is becoming a more cost-effective alternative to acute care for elderly patients in Finland. Yet, daily scheduling of home visits is still frequently done manually, using methods such as spreadsheets, handwritten notes, phone calls, emails, and basic calendars. These processes add to the coordinators' mental load, limit visibility into total capacity and workload balance, result in ineffective routing and uneven workload distribution, and require re-planning in response to cancellations or referrals of urgency, or when staff are unavailable.
The work presented in this thesis described the design and development of a prototype Intelligent Home Hospital Scheduling System (IHHSS) to further HaH coordination by providing integrated information management and transparent decision support. The focus was on reducing the coordinator's burden by automatically proposing feasible, reviewable schedules that consider professional roles and skills, staff availability, and working hours. Patient care can vary depending on patients’ preferred visit times and the feasibility of travel to the clinic, as well as on workload distribution while maintaining the coordinator in control (human-in-the-loop).
A full-stack web application with role-based access was created using React on the frontend and Node.js/Express on the backend, with Supabase (PostgreSQL and auth) for secure storage and role-based access control. The scheduling logic followed a safe-first workflow: unsuitable candidates were filtered using hard constraints (e.g., eligibility and availability), and feasible options were ranked using an interpretable multi-criteria scoring approach (e.g., skill match, availability fit, location, and workload). Time-slot allocation avoids overlapping and observes working-hour limitations. Wherever possible, visit-time preferences were used, and alternatives were offered when necessary. Distance-based sequencing helped with the practical scheduling of visits.
Due to limited access to actual HaH operational data, the thesis used synthetic test scenarios. The findings suggest that the prototype has potential, as it integrates multiple, previously disparate scheduling inputs into a single workflow and generates conflict-free schedules, as demonstrated in experimental scenarios, making replanning faster through reviewable suggestions that coordinators may override. The work has illustrated the potential for transparent, constraint-aware intelligent scheduling support in Finnish HaH services, with advanced optimisation, long-term fairness auditing, and integration into national healthcare systems as future work.
The work presented in this thesis described the design and development of a prototype Intelligent Home Hospital Scheduling System (IHHSS) to further HaH coordination by providing integrated information management and transparent decision support. The focus was on reducing the coordinator's burden by automatically proposing feasible, reviewable schedules that consider professional roles and skills, staff availability, and working hours. Patient care can vary depending on patients’ preferred visit times and the feasibility of travel to the clinic, as well as on workload distribution while maintaining the coordinator in control (human-in-the-loop).
A full-stack web application with role-based access was created using React on the frontend and Node.js/Express on the backend, with Supabase (PostgreSQL and auth) for secure storage and role-based access control. The scheduling logic followed a safe-first workflow: unsuitable candidates were filtered using hard constraints (e.g., eligibility and availability), and feasible options were ranked using an interpretable multi-criteria scoring approach (e.g., skill match, availability fit, location, and workload). Time-slot allocation avoids overlapping and observes working-hour limitations. Wherever possible, visit-time preferences were used, and alternatives were offered when necessary. Distance-based sequencing helped with the practical scheduling of visits.
Due to limited access to actual HaH operational data, the thesis used synthetic test scenarios. The findings suggest that the prototype has potential, as it integrates multiple, previously disparate scheduling inputs into a single workflow and generates conflict-free schedules, as demonstrated in experimental scenarios, making replanning faster through reviewable suggestions that coordinators may override. The work has illustrated the potential for transparent, constraint-aware intelligent scheduling support in Finnish HaH services, with advanced optimisation, long-term fairness auditing, and integration into national healthcare systems as future work.